Figure a ModlfM System with Relocated DCa Bottom Analyz*

CASE STUDY No. 2 Flooded Reflux Drum Inerts Venting (Reference 6)

INSTALLATION Absorption-refrigeration gas plant lean oil still, which separates LPG and gasolines as a top product from absorption oil. The still used a "flooded reflux drum" pressure control method (PC No. 1 in Figure 8.3.2). Pressure was controlled by controlling liquid product leaving the drum. Since the reflux drum was full of liquid (or flooded), closing the valve in the product line would back up liquid into the condenser, flood condenser tubes, reduce the rate of condensation and thus raise column pressure. Similarly, opening the product valve would lower liquid level in the condenser and act to decrease column pressure.

PROBLEM Small quantities of C.s and C-s often entered the column and accumulated as vapor in the reflux arum, thus creating a vapor space near the top of the drum. When the vapor space was formed, column pressure could not be controlled. Normally, manual venting would have been satisfactory to overcome the problem, but the plant was manned only eight hours per day and was operated 24 hours a day.

SOLUTION The automatic venting system shown in Figure 8.3.2 was devised. A second pressure controller (PC No. 2), a level controller and a control valve in the vent line were installed. The set point of PC No. 2 was lower than the set point of the normal pressure controller, PC No. 1. Normally, PC No. 2 was tripped off and did not operate, so that the vent valve was closed and PC No. 1 carried out the control action.

When inerts accumulated, and a vapor gap was formed, the level controller sensed a drop in level. The level controller then sent an air signal that activated PC No. 2. PC No. 2 had a lower set point than PC No. 1 and therefore acted to open the vent valve. As the pressure would fall, PC No. 1 would close, thus helping liquid-level build up. As soon as the inerts were vented and the liquid refilled the drum, the level controller stopped the air signal to PC No. 2, the vent valve closed, and the operation returned to normal.

Carbonate Regenerator

Figur« ¿^Automatic Venting of Hooded Reflux Drum

CASE STUDY No. 3 Condensation in the Vapor Overhead Product Line

(Reference 6)

INSTALLATION A potassium carbonate regenerator in a Benfield gas treating plant (Figure 8.3.3).

PROBLEM Column pressure varied erratically, and could not be controlled properly. In addition, slugs of water entered the sulfur plant.

CAUSE The overhead vapor product pipe, after leaving the reflux drum, was lowered down to grade, and then climbed up to pipe rack level. The back pressure valve was installed at grade to meet the operator's requirement that all control valves be serviceable at grade. This created a low leg in the line. Water due to entrainment and atmospheric condensation accumulated in this leg, and created a significant back pressure, which interfered with the pressure control loop.

CURE The back pressure valve was installed in the pipe rack, and the overhead product pipe was run directly from the reflux drum to the pipe rack. This eliminated the problem.

Temperature Control Loop For Water

CASE STUDY No. 4 Temperature Control in Azeotrope Distillation

(Reference 10)

INSTALLATION A mixed alcohol column (Figure 8.3.4A) separates water and heavy ethers from a mixed alcohol stream. Overhead product is a volatile water-ether azeotrope and a volatile ether-alcohol azeoptrope.

If the quantity of water entering the column is too low, or if excessive heat is supplied to the column (all the water is driven overheads), the azeotropes may not form, causing the heavy ether to pass out of the bottom. To avoid this condition, stripping steam is added to the column.

The control system (Figures 8.3.4A) is an indirect material balance control system. The column base temperature T- was used to regulate the reboil rate. In addition, an intermediate tray temperature T2 regulated the stripping steam input. A rise in the intermediate trav temperature would indicate "drying out" and a rise in the concentration of the heavy ethers; when this occured, steam was injected to re-form the azeotrope.

PROBLEM Excessive losses of alcohol from the still overhead. ANALYSIS Two limitations of the control system were recognized.

1 The base temperature, T,, was not sensitive to small composition changes, and was ineffective in controlling reboil. The operators therefore tended to overreboil the column.

2 Stripping steam was brought on for reasons other than the still drying out. For instance, if feed rate was reduced without reducing the reboil to suit, temperature T2 would rise and bring in stripping steam. A similar effect also followed composition changes.

INVESTIGATION To find the best reboil tray, a temperature profile was taken. It was found that temperature was best for reboil control. Further study, however, showed that variations in the ratio of alcohols present had a much greater effect on the temperature T^ than variations in water concentration. Consequently, this temperatura could not be used for composition control.

It became apparent that a direct measurement of the water content was needed.

SOLUTION A continuous infra-red water analyzer was monitoring for water breakthrough in the bottom stream. This analyzer was modified to sample a plate near the base of the still for water content, and this analysis was used for regulating the reboil rate (Figure 8.3.4B). The modified system lowered alcohol losses by 80 percent, and led to a steadier operation and less operator intervention.

0 0

Post a comment